Localizing images by visual information is a very challenging task in image-based travel recommendations. Travelers take a large number of pictures every day and share them on social networks (Facebook, Sina Weibo, Yelp, etc.). Many of these images are associated with the location where they are taken. But for images that do not associate with geographic location information, how to estimate where they are taken? With the rapid development of social media, the increasing number of shared geographic-labeled images brings an opportunity to address this problem. Using geographic-labeled images to estimate the location of unlabeled images is a popular approach. In this paper, we propose an image geographic location estimation model via multi-task learning (GLML). It combines the classification task and retrieval task to calculate the similarity between the query image and dataset images. Additionally, it fuses multi-global features through multiple global pooling techniques to enhance feature extraction. Each part of the proposed GLML model is flexible and extensible. Experiments on seven public datasets show the effectiveness of the proposed model.
Mangroves are unique forest communities with an abundance of species, high productivity and high ecological, social and economic value. Evaluation of the stress resistance of mangrove plants has mainly focused on the effects of high salinity, heavy metals and flooding, with fewer studies evaluating resistance to upwelling stress. Mangrove species of Avicennia marina, Aegiceras corniculatum and Kandelia obovata were submitted to three temperature upwelling (5, 10 and 15°C) and several physiological and biochemical parameters were measured at six time points (0, 6, 12, 24, 72 and 168 h). The data demonstrated: a certain amount of damage occurred to mangrove plants in the face of prolonged upwelling; different mangrove plants have different response strategies to upwelling; mangrove plants are not sensitive to different upwelling temperatures; the resistance of mangrove plants to upwelling stress was in the following order: A. marina< K. obovata< A. corniculatum. Markers of damage such as relative electrical conductivity (REC), malondialdehyde (MDA) and reactive oxygen species (ROS) among all mangrove species were significantly higher with prolonged upwelling stress. The contents of photosynthetic pigments in all three mangrove species also increased. Superoxide dismutase activity (SOD) was maintained at a high level in both control and treatment groupss. By contrary, the change of peroxidase activity (POD) of A. marina and K. obovata was larger than that of A. corniculatum. Catalase activity (CAT) in A. marina and K. obovata significantly increased under upwelling at both 5 and 10°C while there was no obvious variation of CAT in A. corniculatum. Soluble protein and Soluble sugar contents showed no clear variation but stayed at fairly high levels. However, proline content in A. corniculatum significantly increased under long-term upwelling stress while this was not the case in the other two species. High correlation could be observed between A. marina and MDA, O2- and POD in PCA while A. corniculatum showed association with proline and soluble sugar. In conclusion, the ability of A. corniculatum to tolerate upwelling stress might be due mainly to increases in the activities of SOD and the inducing of proline biosynthesis, while, A marina and K. obovata tolerated upwelling stress by adjusting activity levels of SOD, POD and CAT. Segregation in both principal component analysis (PCA) and hierarchical cluster analysis (HCA) further indicated different tolerances and resistances to upwelling between the three species. Our study provides new insights into the stress response of mangroves to upwelling.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.